owner stringclasses 14
values | repo stringclasses 14
values | id int64 116k 4.61B | issue_number int32 1 181k | author stringlengths 1 39 | body stringlengths 1 262k | created_at timestamp[us]date 2000-06-06 02:40:44 2026-06-02 20:16:18 | updated_at timestamp[us]date 2000-06-06 02:40:44 2026-06-02 20:22:56 | reactions unknown | author_association stringclasses 5
values |
|---|---|---|---|---|---|---|---|---|---|
ClickHouse | ClickHouse | 821,812,755 | 23,199 | Jokser | @taiwhite
Could you please provide, how many parts / partitions are allocated for `s3` table in the case of 10k and 20k rows?
Am I right that the query with `NOT ignore(*)` technically is equal to `SELECT count(*) FROM (SELECT * from s3)` ? | 2021-04-17T12:07:16 | 2021-04-17T12:07:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,819,362 | 23,209 | alexey-milovidov | Race condition in `zlib-ng` on functable - fixed in master. | 2021-04-17T13:02:47 | 2021-04-17T13:02:47 | {} | MEMBER |
ClickHouse | ClickHouse | 821,823,810 | 23,212 | azat | >[Functional stateless tests flaky check (address) — fail: 1, passed: 199](https://clickhouse-test-reports.s3.yandex.net/23212/d02eb3f6b7a8dc07bbed5605e1afcf7634436b3a/functional_stateless_tests_flaky_check_(address).html#fail1)
Will mark `01818_min_bytes_to_use_mmap_io_memory_tracking` as long if the idea of the pa... | 2021-04-17T13:36:28 | 2021-04-17T13:36:28 | {} | MEMBER |
ClickHouse | ClickHouse | 821,841,075 | 23,225 | alexey-milovidov | There are only column comments. Table comments are not implemented. | 2021-04-17T15:31:41 | 2021-04-17T15:31:41 | {} | MEMBER |
ClickHouse | ClickHouse | 821,858,353 | 15,165 | vitlibar | > @MyroTk You have to setup access rights to the Distributed table.
> Internal SELECT or INSERT subqueries that Distributed table is generating are not subjects of access control.
Yes, but it's not how it works right now. If a source table used for a distributed table is located locally, selecting from a distribute... | 2021-04-17T17:32:03 | 2021-04-17T17:32:03 | {} | MEMBER |
ClickHouse | ClickHouse | 821,887,425 | 23,038 | alesapin | > `01305_replica_create_drop_zookeeper` failures.... it can be related to changes...
Got it, we have a bug in the atomicity of drop/create for ReplicatedMergeTree. | 2021-04-17T21:04:50 | 2021-04-17T21:06:06 | {} | MEMBER |
ClickHouse | ClickHouse | 821,974,609 | 23,257 | azat | >[Functional stateless tests (release) — fail: 2, passed: 2984](https://clickhouse-test-reports.s3.yandex.net/23257/0d0e201131df0997e3af66fc7643714f501cb70b/functional_stateless_tests_(release).html#fail1)
Right now it fails because it waits for system log tables | 2021-04-18T11:11:40 | 2021-04-18T11:11:40 | {} | MEMBER |
ClickHouse | ClickHouse | 821,724,837 | 15,101 | alexey-milovidov | @KochetovNicolai it still does not work. | 2021-04-16T23:41:39 | 2021-04-16T23:41:39 | {} | MEMBER |
ClickHouse | ClickHouse | 821,729,063 | 10,571 | alexey-milovidov | Fixed long time ago. | 2021-04-17T00:01:38 | 2021-04-17T00:01:38 | {} | MEMBER |
ClickHouse | ClickHouse | 821,756,215 | 23,105 | 744570676 | @filimonov @KochetovNicolai thank you for your answer. The above result is in version v20.8.9.6-stable, so i try it again in the latest version v21.4.3.21-stable, the query plan is as follows:
```
Expression ((Projection + Before ORDER BY))
Creating... | 2021-04-17T03:07:07 | 2021-04-17T03:07:07 | {} | NONE |
ClickHouse | ClickHouse | 821,772,133 | 21,167 | filimonov | #20547 was merged before forking 21.3.
So all 21.3 versions should include it. Or we know some other issue in that piece of code fixed later @alexey-milovidov? | 2021-04-17T05:49:09 | 2021-04-17T05:49:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,867,403 | 9,363 | alexey-milovidov | The fix is going to be released in version 21.5. | 2021-04-17T18:31:27 | 2021-04-17T18:31:27 | {} | MEMBER |
ClickHouse | ClickHouse | 821,867,438 | 9,363 | alexey-milovidov | 21.5-testing is already available in repositories. | 2021-04-17T18:31:46 | 2021-04-17T18:31:46 | {} | MEMBER |
ClickHouse | ClickHouse | 821,872,001 | 23,220 | vdimir | See PR https://github.com/ClickHouse/ClickHouse/pull/23237 | 2021-04-17T19:05:23 | 2021-04-17T19:05:40 | {} | MEMBER |
ClickHouse | ClickHouse | 821,875,659 | 14,566 | vitlibar | https://github.com/ClickHouse/ClickHouse/pull/23238 | 2021-04-17T19:32:22 | 2021-04-17T19:32:28 | {} | MEMBER |
ClickHouse | ClickHouse | 821,904,634 | 19,816 | nikitamikhaylov | > @nikitamikhaylov Hi! Does this really needs to be documented?
>
> See https://clickhouse.tech/docs/en/sql-reference/functions/ext-dict-functions/#dicthas
>
> No list of supported dictionaries there.
Let's ask @kitaisreal | 2021-04-17T23:58:36 | 2021-04-17T23:58:36 | {} | MEMBER |
ClickHouse | ClickHouse | 821,981,707 | 23,259 | alexey-milovidov | For what reasons the user may want to enable this option?
Maybe better to throw exception that this option is unsupported?
We don't encourage users for fine tuning obscure configuration options. If user makes obscure tuning, it's a bad sign.
Also workarounds on obscure options pollute codebase. | 2021-04-18T12:09:27 | 2021-04-18T12:09:54 | {} | MEMBER |
ClickHouse | ClickHouse | 821,727,017 | 23,202 | nikitamikhaylov | @Mergifyio rebase | 2021-04-16T23:52:09 | 2021-04-16T23:52:09 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 821,728,029 | 13,478 | alexey-milovidov | @Enmk You cannot insert binary data into TEXT column in MySQL.
TEXT column in MySQL always have an encoding and it contains text in this encoding.
In your example the encoding is unspecified. It can be UTF-8 if you are lucky.
But as the issue exists - most likely it is `latin1` (with Swedish collation).
So you just... | 2021-04-16T23:56:58 | 2021-04-16T23:57:47 | {} | MEMBER |
ClickHouse | ClickHouse | 821,821,090 | 23,212 | azat | >In my estimate it should shine on large reads.
Agree, there are use cases when this should work well, something like reading all data from the one table in multiple queries, but how much better it will? Do you have any numbers? | 2021-04-17T13:15:45 | 2021-04-17T13:15:45 | {} | MEMBER |
ClickHouse | ClickHouse | 821,859,646 | 14,810 | vitlibar | Inserting into the source table of a material view causes automatically inserting into the materialized view's target table.
Both operations are executing in the same context, that's why you should have the INSERT privilege on both the source table and the target table to make the inserting into the source table.
I... | 2021-04-17T17:41:56 | 2021-04-17T17:41:56 | {
"+1": 2
} | MEMBER |
ClickHouse | ClickHouse | 821,973,043 | 19,320 | pavelnemirovsky | +1 | 2021-04-18T10:59:25 | 2021-04-18T10:59:25 | {} | NONE |
ClickHouse | ClickHouse | 821,732,457 | 6,972 | alexey-milovidov | No info provided. | 2021-04-17T00:15:31 | 2021-04-17T00:15:31 | {} | MEMBER |
ClickHouse | ClickHouse | 821,771,541 | 21,167 | filimonov | Hm. Didn't check it but thought that all 21.3 have related patches.
Let's redirect that question to @chyy - can you please try the 21.3.7? | 2021-04-17T05:42:43 | 2021-04-17T05:43:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,811,927 | 15,165 | filimonov | > Internal SELECT or INSERT subqueries that Distributed table is generating are not subjects of access control.
Selects are. Inserts - not sure how it works with `secret`. Do you know it @azat? | 2021-04-17T12:00:12 | 2021-04-17T12:00:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,822,523 | 14,999 | alexey-milovidov | The original intent of the developer was to count time from the insertion time, not from real time
\- to keep more nodes instead of less.
Actually the intent is to keep as many deduplication nodes as possible (so ideally, deduplication will work forever) and the setting exists only to avoid using too many nodes in ... | 2021-04-17T13:26:54 | 2021-04-17T13:28:09 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 821,904,107 | 19,816 | olgarev | @nikitamikhaylov Hi! Does this really needs to be documented?
See https://clickhouse.tech/docs/en/sql-reference/functions/ext-dict-functions/#dicthas
No list of supported dictionaries there. | 2021-04-17T23:52:51 | 2021-04-17T23:52:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,942,015 | 23,245 | ucasfl | I can't reproduce this in my environment with the latest master branch. | 2021-04-18T06:26:49 | 2021-04-18T06:26:49 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 821,728,587 | 11,999 | alexey-milovidov | @Enmk Looks like we have to do it on our own. | 2021-04-16T23:59:44 | 2021-04-16T23:59:44 | {} | MEMBER |
ClickHouse | ClickHouse | 821,733,802 | 7,701 | CurtizJ | In modern ClickHouse versions (20.8+) TTL should work fine. | 2021-04-17T00:21:37 | 2021-04-17T00:21:37 | {} | MEMBER |
ClickHouse | ClickHouse | 821,825,886 | 23,038 | alesapin | So ZooKeeper still 2-3 times better for write requests.
BTW results with three nodes look quite strange to me (all requests too slow). | 2021-04-17T13:51:54 | 2021-04-17T13:56:03 | {
"confused": 1
} | MEMBER |
ClickHouse | ClickHouse | 821,723,896 | 14,511 | alexey-milovidov | There is a chance for user to make misconfiguration - minor usability issue. | 2021-04-16T23:37:44 | 2021-04-16T23:37:44 | {} | MEMBER |
ClickHouse | ClickHouse | 821,734,250 | 18,151 | alexey-milovidov | @zhang2014 Your help will be really appreciated. | 2021-04-17T00:23:51 | 2021-04-17T00:23:51 | {} | MEMBER |
ClickHouse | ClickHouse | 821,765,048 | 5,429 | azat | @alexey-milovidov list of non-materialized columns is obtained from the Distributed table, while in #16897 the Distributed table has all columns non-materialized that's why it wasn't filtered out and rejected. | 2021-04-17T04:39:14 | 2021-04-17T04:39:14 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 821,818,429 | 23,212 | alexey-milovidov | In my estimate it should shine on large reads. | 2021-04-17T12:54:47 | 2021-04-17T12:54:47 | {} | MEMBER |
ClickHouse | ClickHouse | 821,818,474 | 23,212 | alexey-milovidov | See the MSan report in `OptimizeShardingKeyRewriteInVisitor.cpp:50`. | 2021-04-17T12:55:11 | 2021-04-17T12:55:11 | {} | MEMBER |
ClickHouse | ClickHouse | 821,824,640 | 21,639 | alexey-milovidov | With Ryzen 3950 CPU you can have compilation time around 10 minutes, and the total cost of the workstation will be slightly less than MacBook laptop. | 2021-04-17T13:42:55 | 2021-04-17T13:44:18 | {
"+1": 1,
"-1": 3
} | MEMBER |
ClickHouse | ClickHouse | 821,827,194 | 23,212 | alexey-milovidov | > but how much better it will? Do you have any numbers?
Yes, it's in perf tests in a PR where it was enabled. | 2021-04-17T14:00:15 | 2021-04-17T14:00:23 | {} | MEMBER |
ClickHouse | ClickHouse | 821,838,682 | 14,511 | traceon | Unverified effects of role adjustments may easily introduce security breaches in this case. | 2021-04-17T15:15:35 | 2021-04-17T15:15:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,866,834 | 20,216 | alexey-milovidov | So, if you can construct binary format of aggregate function state by hand and pass it to ClickHouse - it is fine.
I don't know any libraries that are doing that. | 2021-04-17T18:28:06 | 2021-04-17T18:28:06 | {} | MEMBER |
ClickHouse | ClickHouse | 821,885,168 | 22,237 | alexey-milovidov | SQLancer: https://github.com/ClickHouse/ClickHouse/issues/23029 | 2021-04-17T20:47:01 | 2021-04-17T20:47:01 | {} | MEMBER |
ClickHouse | ClickHouse | 821,952,000 | 15,972 | francisoliverlee | > I can successfully build clickhouse when i modify Sources.h file, but i am not sure if there is a potential risk with this patch.
>
> ```
> diff --git a/src/Functions/GatherUtils/Sources.h b/src/Functions/GatherUtils/Sources.h
> index 299884e1c9..6a0c718b02 100644
> --- a/src/Functions/GatherUtils/Sources.h
> ... | 2021-04-18T08:05:43 | 2021-04-18T08:05:43 | {} | NONE |
ClickHouse | ClickHouse | 821,719,927 | 14,610 | alexey-milovidov | It's fixed:
```
$ clickhouse-client
ClickHouse client version 21.5.1.1.
Connecting to localhost:9000 as user default.
Connected to ClickHouse server version 21.5.1 revision 54448.
milovidov-desktop :) select case 1.1 when 0.1 then 'a' when 1.1 then 'b' when 2.1 then 'c' else 'default' end as f
SELECT case... | 2021-04-16T23:20:56 | 2021-04-16T23:20:56 | {} | MEMBER |
ClickHouse | ClickHouse | 821,723,166 | 14,566 | alexey-milovidov | @vitlibar Said that it works as expected, but documentation can be misleading. | 2021-04-16T23:34:41 | 2021-04-16T23:34:41 | {} | MEMBER |
ClickHouse | ClickHouse | 821,723,283 | 14,476 | alexey-milovidov | `SELECT CAST([] AS Array(Array(String)))` | 2021-04-16T23:35:05 | 2021-04-16T23:35:05 | {} | MEMBER |
ClickHouse | ClickHouse | 821,733,360 | 17,207 | alexey-milovidov | @ithangzhou It may happen if you created an instance with low amount of memory (4 GB or lower).
In this case the server may not being able to send an exception about not enough memory and closing a connection.
Check server logs. | 2021-04-17T00:19:32 | 2021-04-17T00:19:32 | {} | MEMBER |
ClickHouse | ClickHouse | 821,817,184 | 21,639 | traceon | @nvartolomei
MacBook Air M1, 8 CPU + 8 GPU, 16 GB RAM
---
AppleClang + Ninja toolchain:
`cmake -DCMAKE_BUILD_TYPE=RelWithDebInfo ..`
14,52s user 8,40s system 77% cpu 29,555 total
`cmake --build . --config RelWithDebInfo --parallel`
9417,54s user 503,73s system 537% cpu 30:46,89 total
---
AppleC... | 2021-04-17T12:45:16 | 2021-04-17T12:45:36 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,866,609 | 20,216 | alexey-milovidov | Sometimes we have to change the internal representation of aggregation states.
We are going to add some means for compatibility: https://github.com/ClickHouse/ClickHouse/issues/12552
But it will be only one way compatibility: ClickHouse will be able to read older states of aggregate functions. But it will not seria... | 2021-04-17T18:26:49 | 2021-04-17T18:26:49 | {} | MEMBER |
ClickHouse | ClickHouse | 821,823,670 | 15,101 | vdimir | > LowCardinality of Nullable is Ok.
@alexey-milovidov
~~Now `canBeInsideNullable` is `false` for `LowCardinality(String)`. Enabling it in `ColumnLowCardinality.h` and `DataTypeLowCardinality.h` makes this query work, I'll watch on other tests in CI #23220~~
**UPD:** this is about `Nullable(LowCardinality(T))`... | 2021-04-17T13:35:35 | 2021-04-17T19:07:42 | {} | MEMBER |
ClickHouse | ClickHouse | 821,879,314 | 23,038 | alesapin | `01305_replica_create_drop_zookeeper` failures.... it can be related to changes... | 2021-04-17T20:00:37 | 2021-04-17T20:00:37 | {} | MEMBER |
ClickHouse | ClickHouse | 821,725,383 | 21,639 | kitaisreal | @changvvb, @traceon thanks for the job. Great contribution. | 2021-04-16T23:44:12 | 2021-04-16T23:44:12 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,769,910 | 23,174 | azat | > if we bind to one numa
percpu_arena with limited CPUs to run on has this known issue, but it is not the case for clickhouse since it uses patched version of jemalloc, that as the following patch - 93e27e435cac846028da20cd9b0841fbc9110bd2, which basically ignore CPU limiting and will always use narenas=all-availabl... | 2021-04-17T05:26:32 | 2021-04-17T05:26:32 | {} | MEMBER |
ClickHouse | ClickHouse | 821,795,053 | 18,213 | tavplubix | Duplicate of #18926, fixed in #20878 | 2021-04-17T09:21:49 | 2021-04-17T09:21:49 | {} | MEMBER |
ClickHouse | ClickHouse | 821,824,253 | 16,897 | alexey-milovidov | Another clarification from @azat:
\- you can make column MATERIALIZED in the Distributed table identically to the local table - then everything will work as expected. | 2021-04-17T13:39:54 | 2021-04-17T13:39:54 | {} | MEMBER |
ClickHouse | ClickHouse | 821,855,884 | 23,054 | kozlek | Just updated, it's working perfectly 👌
Thanks a lot 🙏 | 2021-04-17T17:13:39 | 2021-04-17T17:13:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,982,622 | 23,261 | alexey-milovidov | The settings won't be enabled until we recieve an answer from @KochetovNicolai here: https://github.com/ClickHouse/ClickHouse/issues/21167#issuecomment-820907631
and the potential issue found here #23257 will be investigated.
CC @KochetovNicolai @Avogar | 2021-04-18T12:16:45 | 2021-04-18T12:16:45 | {} | MEMBER |
ClickHouse | ClickHouse | 821,821,705 | 22,242 | kitaisreal | @hexiaoting can this pull request be reviewed ? Or it is still draft ? | 2021-04-17T13:20:36 | 2021-04-17T13:20:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,868,544 | 23,183 | alexey-milovidov | @roygbip Could you help me debugging this issue? It happens on specific cloud instance and on specific OS where I cannot obtain access. Maybe you can create a small temporary VM, temporary ssh keypair and send the address along with keypair to clickhouse-feedback@yandex-team.ru? Then we will need a few days to experime... | 2021-04-17T18:41:02 | 2021-04-17T18:41:10 | {} | MEMBER |
ClickHouse | ClickHouse | 821,868,932 | 23,050 | alexey-milovidov | @Mergifyio rebase | 2021-04-17T18:44:07 | 2021-04-17T18:44:07 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 821,982,018 | 23,245 | alexey-milovidov | @ucasFL Did you upload the data to `test.hits` table? | 2021-04-18T12:12:08 | 2021-04-18T12:12:08 | {} | MEMBER |
ClickHouse | ClickHouse | 821,731,622 | 7,701 | alexey-milovidov | @CurtizJ did not answer, so the issue is fixed for sure. | 2021-04-17T00:11:57 | 2021-04-17T00:11:57 | {} | MEMBER |
ClickHouse | ClickHouse | 821,820,162 | 22,237 | azat | >Functional stateless tests (thread) — fail: 5, passed: 478
#23153
>AST fuzzer (ASan) — Fuzzer failed (1). See the logs.
SIGKILL
>Integration tests (asan) — fail: 26, passed: 777, flaky: 5
docker-compose pull fails even after retries | 2021-04-17T13:09:11 | 2021-04-17T13:09:11 | {} | MEMBER |
ClickHouse | ClickHouse | 821,820,501 | 23,190 | alexey-milovidov | Need to check `test_keeper_multinode_blocade_leader`. | 2021-04-17T13:11:44 | 2021-04-17T13:11:44 | {} | MEMBER |
ClickHouse | ClickHouse | 821,822,861 | 14,999 | alexey-milovidov | @FLPhoenix
> On the next try, the same software drops previously inserted rows
Do you mean dropping rows by ALTER DELETE? | 2021-04-17T13:29:36 | 2021-04-17T13:29:58 | {} | MEMBER |
ClickHouse | ClickHouse | 821,823,787 | 21,904 | tlightsky | I got same issue building on osx,
both for gcc-10 and clang-11
```
clang version 11.1.0
Target: x86_64-apple-darwin20.3.0
Thread model: posix
InstalledDir: /usr/local/opt/llvm/bin
```
```
ClickHouse/contrib/libcxx/include/cstdlib:158:9: error: no member named 'aligned_alloc' in the global namespace
using :... | 2021-04-17T13:36:21 | 2021-04-17T13:36:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,825,785 | 23,038 | alesapin | Three nodes:
### Create requests with 32 bytes payload (30 threads):
Keeper + fsync:
```
read requests 0, write requests 10443, Write RPS: 613.018, Write MiB/s: 0.0894468.
Write sampler:
0% 0.0136349 sec.
10% 0.031579 sec.
20% 0.0368288 sec.
30% 0.0417666 s... | 2021-04-17T13:51:18 | 2021-04-17T13:59:15 | {} | MEMBER |
ClickHouse | ClickHouse | 821,853,123 | 3,483 | Tasselmi | Hey, do you want some help？ actually I solved this problem by myself | 2021-04-17T16:54:27 | 2021-04-17T16:54:27 | {} | NONE |
ClickHouse | ClickHouse | 821,869,055 | 23,050 | mergify[bot] | **Command `rebase`: success**
> **Branch has been successfully rebased**
>
| 2021-04-17T18:45:14 | 2021-04-17T18:45:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,816,518 | 21,639 | nvartolomei | @changvvb can you share some numbers for compilation durations? Wondering if it is practical to use an M1 laptop as a development machine for ClickHouse. | 2021-04-17T12:39:21 | 2021-04-17T12:39:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,840,893 | 14,511 | alexey-milovidov | @traceon Do you want to take this task? | 2021-04-17T15:30:24 | 2021-04-17T15:30:24 | {} | MEMBER |
ClickHouse | ClickHouse | 821,730,014 | 10,019 | alexey-milovidov | @filimonov See #19576. | 2021-04-17T00:05:04 | 2021-04-17T00:05:04 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 821,737,057 | 17,673 | alexey-milovidov | See https://github.com/ClickHouse/ClickHouse/pull/11120 | 2021-04-17T00:39:43 | 2021-04-17T00:39:43 | {} | MEMBER |
ClickHouse | ClickHouse | 821,839,441 | 21,904 | traceon | @tlightsky make sure you clone the recent `master` recursively with submodules, verify with `git status` that everything is correct. Update your macOS and Xcode.
Precisely follow the official macOS build instructions: https://clickhouse.tech/docs/en/development/build-osx/ | 2021-04-17T15:20:50 | 2021-04-17T15:20:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,726,816 | 13,478 | alexey-milovidov | Fixed here: #19800. | 2021-04-16T23:51:14 | 2021-04-16T23:51:34 | {} | MEMBER |
ClickHouse | ClickHouse | 821,735,476 | 19,335 | alexey-milovidov | No reports in past three months. | 2021-04-17T00:30:41 | 2021-04-17T00:30:41 | {} | MEMBER |
ClickHouse | ClickHouse | 821,818,285 | 23,211 | alexey-milovidov | > RSS will grow, total_memory_tracker will be synced with RSS and the allocations will start to fail.
But how does it help to add something to memory tracker?
The total memory tracker will still be corrected to higher value. | 2021-04-17T12:53:35 | 2021-04-17T12:53:35 | {} | MEMBER |
ClickHouse | ClickHouse | 821,819,219 | 23,038 | alesapin | Wow, nothing is broken. | 2021-04-17T13:01:49 | 2021-04-17T13:01:49 | {} | MEMBER |
ClickHouse | ClickHouse | 821,823,531 | 23,212 | azat | >See the MSan report in OptimizeShardingKeyRewriteInVisitor.cpp:50.
>[AST fuzzer (MSan) — SUMMARY: MemorySanitizer: use-of-uninitialized-value obj-x86_64-linux-gnu/../src/Interpreters/OptimizeShardingKeyRewriteInVisitor.cpp:50:28](https://clickhouse-test-reports.s3.yandex.net/23212/d02eb3f6b7a8dc07bbed5605e1afcf763443... | 2021-04-17T13:34:39 | 2021-04-17T13:34:39 | {} | MEMBER |
ClickHouse | ClickHouse | 821,827,337 | 23,212 | alexey-milovidov | #22206 #22326 | 2021-04-17T14:01:18 | 2021-04-17T14:01:18 | {} | MEMBER |
ClickHouse | ClickHouse | 821,832,934 | 23,214 | den-crane | You can use:
```sql
select A.country, B.area from city A inner join (select * from city where area='south') B on A.city=B.city
```
This is actually CH conventional way, because 3 years ago ON syntax was not implemented yet. And Clickhouse JOIN-rewriter transforms queries with join into this form with subqueries. | 2021-04-17T14:39:59 | 2021-04-17T14:39:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,835,901 | 21,639 | traceon | @alexey-milovidov I doubt that anyone buying a MacBook is optimizing for cpu performance-to-cost only.
Ryzen 3950x CPU only is around $725. How can you build a workstation around it for 1200 - 725 = $475 ?? You are going to spend half of that remaining money on a power unit only. | 2021-04-17T14:58:18 | 2021-04-17T15:07:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,897,240 | 23,229 | sevirov | Internal documentation ticket: DOCSUP-8654. | 2021-04-17T22:40:04 | 2021-04-17T22:40:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,936,468 | 23,212 | azat | >In my estimate it should shine on large reads.
I'm fine with changing the default to any sensible value (1-10GB), but after #23211 it is better to have max limit for mmap.
>#22206
Seems that hits_none created from hits (in 00162_mmap_compression_none.sql) and hits that I have is:
```
65M hits/data/dat... | 2021-04-18T05:27:29 | 2021-04-18T05:27:29 | {} | MEMBER |
ClickHouse | ClickHouse | 821,730,292 | 9,932 | alexey-milovidov | Need info from @filimonov | 2021-04-17T00:06:08 | 2021-04-17T00:06:08 | {} | MEMBER |
ClickHouse | ClickHouse | 821,737,951 | 22,410 | ithangzhou | when clickhouse execute the distrubite-ddl,it will put the task to task-queue in zk-cluster,so other clickhouse server can fetch the submitted task from the queue and redo it.If your zk-cluster has low performance with IO processing,it may cause task delay for a long period of time.So,try to upgrade your zk-cluster wit... | 2021-04-17T00:45:42 | 2021-04-17T00:45:42 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,818,352 | 21,639 | changvvb | > @changvvb can you share some numbers for compilation durations?
@nvartolomei It spends about 20 - 30 minutes to compile the entire project, faster than X86_64. | 2021-04-17T12:54:10 | 2021-04-17T12:54:10 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,916,588 | 22,984 | benbiti | merge master and update | 2021-04-18T01:56:23 | 2021-04-18T01:56:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,947,526 | 19,816 | kitaisreal | @olgarev no this feature is not needed to be documented. | 2021-04-18T07:25:24 | 2021-04-18T07:25:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,820,760 | 23,211 | azat | > The total memory tracker will still be corrected to higher value.
Right, but this is not enough
> But how does it help to add something to memory tracker?
Suppose you have lots of small queries and few big queries that just read all the data from the table, in this case one of that big queries will mmap all ... | 2021-04-17T13:13:23 | 2021-04-17T13:13:23 | {} | MEMBER |
ClickHouse | ClickHouse | 821,844,783 | 14,511 | traceon | @alexey-milovidov Nope :) | 2021-04-17T15:57:15 | 2021-04-17T15:57:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,908,450 | 21,589 | alexey-milovidov | @akuzm This PR is unfinished. | 2021-04-18T00:34:02 | 2021-04-18T00:34:02 | {} | MEMBER |
ClickHouse | ClickHouse | 821,909,283 | 21,589 | alexey-milovidov | This function should be reimplemented in vectorized fashion, as we discussed.
1. Filter arrays that have at least N (=1) elements.
2. Extract the values of the Nth elements of every array.
3. Extract the values of accumulator at the corresponding positions.
4. Apply lambda function and obtain the new values of th... | 2021-04-18T00:43:36 | 2021-04-18T00:46:58 | {} | MEMBER |
ClickHouse | ClickHouse | 821,980,707 | 23,258 | alexey-milovidov | @azat We build our sources with `-Werror`.
All builds in CI are using `-Werror` with the only exception of "unbundled" build. | 2021-04-18T12:02:13 | 2021-04-18T12:05:23 | {} | MEMBER |
ClickHouse | ClickHouse | 821,731,870 | 7,697 | alexey-milovidov | @filimonov restrained to either accept or close this issue. | 2021-04-17T00:12:59 | 2021-04-17T00:12:59 | {} | MEMBER |
ClickHouse | ClickHouse | 821,908,881 | 21,589 | alexey-milovidov | @akuzm Please improve performance at least 1000 times. | 2021-04-18T00:39:03 | 2021-04-18T00:39:03 | {} | MEMBER |
ClickHouse | ClickHouse | 821,980,873 | 23,258 | alexey-milovidov | ```
grep -F -- '-Werror' compile_commands.json | wc -l
9640
```
| 2021-04-18T12:03:22 | 2021-04-18T12:03:22 | {} | MEMBER |
ClickHouse | ClickHouse | 815,081,232 | 17,708 | alexey-milovidov | @nrolf We will be happy to help.
https://clickhouse.tech/docs/en/development/developer-instruction/ | 2021-04-07T17:12:22 | 2021-04-07T17:12:22 | {} | MEMBER |
ClickHouse | ClickHouse | 815,303,791 | 22,408 | UnamedRus | https://github.com/tdunning/t-digest/commit/72a1d6fc85523b046246fd16bcd3f3df5c689d3b
Before this commit: 7199.999999999989
After this commit: 9000.0
| 2021-04-07T22:29:36 | 2021-04-07T22:29:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,558,668 | 22,827 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/pull/7651 | 2021-04-08T08:19:58 | 2021-04-08T08:19:58 | {} | MEMBER |
ClickHouse | ClickHouse | 815,646,280 | 22,619 | olgarev | > @olgarev @KochetovNicolai You forgot about the bonus:
> [ac3de66](https://github.com/ClickHouse/ClickHouse/commit/ac3de6626af6772297c2f5df653854b98635605d)
@KochetovNicolai The PR was only about dateTime64, not about date or dateTime. | 2021-04-08T10:25:43 | 2021-04-08T10:25:43 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.